Many of us have seen films containing remarkably realistic dinosaurs, aliens, animated toys and other fanciful creatures. Such animations are made possible by computer graphics. Using such techniques, a computer graphics artist can specify how each object should look and how it should change in appearance over time, and a computer then models the objects and displays them on a display such as your television or a computer screen. The computer takes care of performing the many tasks required to make sure that each part of the displayed image is colored and shaped just right based on the position and orientation of each object in a scene, the direction in which light seems to strike each object, the surface texture of each object, and other factors.
Because computer graphics generation is complex, computer-generated three-dimensional graphics just a few years ago were mostly limited to expensive specialized flight simulators, high-end graphics workstations and supercomputers. The public saw some of the images generated by these computer systems in movies and expensive television advertisements, but most of us couldn't actually interact with the computers doing the graphics generation. All this has changed with the availability of relatively inexpensive 3D graphics platforms such as, for example, the Nintendo 64® and various 3D graphics cards now available for personal computers. It is now possible to interact with exciting 3D animations and simulations on relatively inexpensive computer graphics systems in your home or office.
A problem graphics system designers confronted in the past was how to efficiently implement shaders in a graphics system. Generally, shading is the process performing lighting computations and determining pixel colors/opacities from them. Generally, there are three main types of shading in common use: flat, Gouraud, and Phong. These correspond to computing the light per polygon, per vertex and per pixel. A wide variety of shading models have been created. There is no one shading model that pleases all users and is suitable for all applications. Therefore, several design approaches have been suggested to provide flexibility in terms of programmer selection and specification of shading models.
In the paper by R. L. Cook called “Shade Trees” (SIGGRAPH 84, pages 223-231, the author described a special purpose language in which a shader is built as a tree expression called a shade tree. Generally speaking, a shade tree is a tree of nodes each of which takes parameters from its children and produces parameters for its parent. For example, the parameters may be the terms of the illumination equation (e.g., specular coefficient or surface Normal). Other parameters might comprise atmospheric effects (e.g., haze) or projections. The RenderMan Interface uses shade trees to provide user-defined and system-defined shaders for a variety of purposes.
While shade trees have been used extensively in non-real-time rendering graphics systems, problems arise when trying to accommodate the flexibility that shade trees provide within the context of real-time rendering. It would be highly desirable to be able to provide the flexibility of shade trees within low cost real-time rendering systems such as, for example, home video game platforms and personal computer graphics cards.
Another problem confronting graphics systems designers has been how to efficiently provide a feature called single-pass multitexturing. Basically, texturing is a technique for efficiently modeling the properties of a surface. For example, instead of modeling the geometry of each individual brick and mortar line within a brick wall, it is possible to electronically “glue” an image of a brick wall onto a surface. Such texturing capabilities can be used to significantly increase image complexity without a corresponding increase in modeling and processing costs.
The extension to texturing known as multitexturing allows two or more textures to be applied to the same surface. For example, suppose you want to create an image of the earth as it might be seen from outer space. You could model the earth as a sphere and apply two different textures to it. The first texture could be an image of the continents and oceans. The second texture could be an image of cloud cover. By moving the cloud cover texture image relative to the continent/ocean texture image, you could create a very realistic dynamic texture-mapped image.
Some graphics accelerators support multitexturing in which two or more textures are accessed during the same rendering pass. See, for example, Microsoft's Direct X 6.0 SBK (1998); Segal et al., “The Open GL Graphics System: A Specification” (Version 1.2.1) (March 1998) (www.OpenGL.org). Certain PC graphics accelerator cards also provide single pass multitexturing. However, further improvements are possible.
The present invention provides a generalized shade tree blender that can be used for multitexturing as well as a number of other flexible blending effects. In accordance with one aspect provided by this invention, recirculating shader hardware within a graphics pipeline can be controlled to provide a number of independently controllable blending stages. A shader hardware includes intermediate storage for results of previous blending operations. The shader hardware can select different inputs and perform different operations for each blending stage. Thus, relatively low cost and compact shader hardware can be used to implement arbitrarily complex shade trees.
In accordance with another aspect provided by this invention, the results of a first texture mapping operation is provided to a reconfigurable shader. The shader performs a blending operation in response to the first texture mapping operation. The shader is then reconfigured, and is connected to receive the results of a further texturing operation. The reconfigured shader combines its previous results with the results of the further texturing operation to provide a blended output.
In accordance with a further aspect provided by this invention, a shader can be recirculated any desired number of times to implement an arbitrarily complex shading model. Each recirculation or “stage” can be programmed to have any one of a number of desired blending operations and to blend from selected ones of a variety of color, opacity or depth sources. The number of recirculations may be limited in a particular implementation in view of real-time rendering timing constraints, but a reasonable number of recirculation stages (e.g., fifteen) can provide great flexibility in implementing a variety of complex shading models.
In accordance with another aspect provided by this invention, a recirculating shade tree pixel blender is implemented in hardware to minimize processing time per stage. In more detail, a preferred embodiment of this invention provides a relatively low chip-footprint, versatile texture-environment processing subsystem including a hardware accelerated programmable texture shader/pixel blender that circulates computed color, opacity and other data over multiple cycles/stages. The texture environment subsystem can combine per-vertex lighting, textures, rasterized colors, opacities, and depths to form pixel parameters for display. Blending operations for color (e.g., RGB) and alpha components may be independently processed within the texture environment subsystem by a blending unit comprising a set of color/alpha combiner (shader) hardware that is reused over multiple processing stages to implement multitexturing and other effects. Selectable current-color/opacity input/output registers may be shared among all stages to store intermediate results. The shader hardware can be reconfigured for each stage to provide a chain of specifiable blending/shading operations supporting single rendering pass multitexturing and other effects.